Development of a fuzzy-fule-base system with educational applications with case study
In criterion-referenced assessment method (CRA), total score for students’ work is gather by summing up the scores for each main criterion, where total score is overall mark awarded to student for their performed work e.g. assignment, test, project and etc. CRA is a linear assessment method wher...
Saved in:
| 主要作者: | |
|---|---|
| 格式: | Final Year Project Report |
| 語言: | English English |
| 出版: |
Universiti Malaysia Sarawak, UNIMAS
2009
|
| 主題: | |
| 在線閱讀: | http://ir.unimas.my/id/eprint/6509/1/DEVELOPMENT%20OF%20A%20FUZZY-RULE-BASE%20SYSTEM%20WITH%20EDUCATIONAL%20APPLICATIONS%20WITH%20CASE%20STUDY%2824%20pgs%29.pdf http://ir.unimas.my/id/eprint/6509/8/LEE%20KIM%20KHOON.pdf http://ir.unimas.my/id/eprint/6509/ |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | In criterion-referenced assessment method (CRA), total score for students’
work is gather by summing up the scores for each main criterion, where total score
is overall mark awarded to student for their performed work e.g. assignment, test,
project and etc. CRA is a linear assessment method where total score varies in
direct proportion to the scores from each main criterion. A fuzzy inference system
(FIS) based assessment model is proposed and developed to allow non-linear
relationship between total score and score from each main criterion. FIS based
assessment model is constructed with expert knowledge, rules collected from human
expert are stored in fuzzy rule base for the use of inference. The number of rules
increases exponentially as the number of main criteria increase. As a solution to this
issue, a rule reduction system (RRS) is developed. The RRS can pin point a set of
important rules, and it is suggested that only important rules is collected. A case
study is conducted to evaluate the performance of the developed system. Empirical
results show that the FIS based assessment model allow the non-linear relation
among total score and the scores from each main criterion to be modeled. Besides,
experiments show that the developed RRS can reduce the fuzzy rule significantly. |
|---|